Hamstring muscle strain is the most prevalent injury in Australian Rules Football, accounting for 16% of playing time missed as a result of injury. Thirty-seven professional footballers from an Australian Football League team had preseason measurements of hamstring and quadriceps muscle concentric peak torque at 60, 180, and 300 deg/sec measured on a Cybex 340 dynamometer. Players were studied prospectively throughout the 1995 season. During that time, six players sustained clinically diagnosed hamstring muscle injuries that caused them to miss match-playing time. The injured hamstring muscles were all weaker than in the opposite leg in absolute values and hamstring-to-quadriceps muscle ratios. According to our t-test results, hamstring muscle injury was significantly associated with a low hamstring-to-quadriceps muscle peak torque ratio at 60 deg/sec on the injured side and a low hamstring muscle side-to-side peak torque ratio at 60 deg/sec. Flexibility (as measured by the sit-and-reach test) did not correlate with injury. Discriminant-function analysis using the two significant ratio variables resulted in a canonical correlation with injury of 0.4594 and correctly classified legs into injury groups with 77.4% success. These results indicate that preseason isokinetic testing of professional Australian Rules footballers can identify players at risk of developing hamstring muscle strains.